Field-collected house finches of mixed sex and age were infected experimentally with either western equine encephalomyelitis (WEE) or St. Louis encephalitis (SLE) viruses during the summer or fall of 1998 and maintained over the winter under ambient conditions. To detect natural relapse during the spring, 32 birds were bled weekly from February through June 1999, and then necropsied 1 yr after infection to detect chronic infections using a reverse transcription polymerase chain reaction (RT-PCR). After 10 mo, 13/14 surviving birds previously infected with WEE were antibody positive by enzyme immunoassay (EIA), and 11/14 had plaque reduction neutralization test (PRNT) antibody titers >1:20, whereas only of 8/13 birds previously infected with SLE were positive by EIA and all had PRNT titers <1:20. When necropsied, 1/14 and 1/13 birds had WEE and SLE RT-PCR positive lung or spleen tissue, respectively; blood, brain, and liver tissues were negative as were all previous blood samples. All tissues from these birds including weekly blood samples tested negative for infectious virus by plaque assay on Vero cell culture. To determine if persistent antibody was protective, birds infected initially with WEE or SLE in November 1998 were challenged 6 mo later with homologous virus. WEE antibody persisted well (5/6 birds remained PRNT positive before challenge) and remained protective, because 0/6 birds were viremic after challenge. In contrast, SLE antibody decayed rapidly (0/6 birds remained PRNT positive before challenge) and was not protective, because 3/6 birds developed an ephemeral viremia on day 1 after infection (mean titer, 10^2.73 plaque forming units/0.1 ml). When necropsied 7 wk after challenge, 1/10 birds infected with WEE and 1/10 birds infected with SLE exhibited an RT-PCR positive spleen, despite the fact that both birds had PRNT antibody titers >1:40 at this time. To determine if immunosuppression would cause a chronic infection to relapse, eight birds initially infected with either WEE or SLE were treated with cyclophosphamide and then tested repeatedly for viremia; all samples were negative for virus by plaque assay. Collectively, our results indicated that a low percentage of birds experimentally infected with WEE or SLE developed chronic infections in the spleen or lung that could be detected by RT-PCR, but not by plaque assay. Birds did not appear to relapse naturally or after immunosuppression. The rapid decay of SLE, but not WEE, antibody may allow the relapse of chronic infections of SLE, but not WEE, to produce viremias sufficiently elevated to infect mosquitoes.